Mediating Investor Attention - Berkeley Haas · Mediating Investor Attention Abstract We review the...
Transcript of Mediating Investor Attention - Berkeley Haas · Mediating Investor Attention Abstract We review the...
Mediating Investor Attention
Brad M. Barber Graduate School of Management University of California, Davis
Shengle Lin College of Business
San Francisco State University
Terrance Odean Haas School of Business
University of California, Berkeley
(1st Draft)
April 1, 2019 _________________________________
We are grateful to Paul Tetlock for providing news data and to Clifton Green and Yabin Wu for assistance analyzing 2007-2017 TAQ data. We have benefited from the comments of conference participants at the Conference of Research on Collective Goods at the Max Plank Institute in Bonn, the Brazilian Finance Meeting, and the Boulder Summer Conference on Consumer Finance Decision Making.
Mediating Investor Attention Abstract
We review the literature on investor attention with a focus on studies of events that attract
investors’ attention. Such events are associated with sharp short-term price reactions,
often followed by reversals, an asymmetry in price reactions with stronger responses to
positive signals, increases in trading volume, and an asymmetrical effect on the buying
and selling of individual investors who tend to be on the buy side of the market for
attention-attracting stocks. Most studies we discuss document some, but not all, of these
phenomena. We analyze 1983-2000 ISSM/TAQ and 2007-2017 TAQ data to show that,
for several previously studied attention-attracting events, individual investors are on the
buy side of the market. We argue that the primary determinant of individual investor
attention is media coverage and location and we discuss support in the literature for this
hypothesis.
2
When faced with a large number of choices, how people allocate attention may
determine their choices as much or more than preferences and beliefs. For example,
consider an individual investor choosing a U.S. listed stock to purchase. She faces
thousands of alternatives. It is the rare investor indeed who will carefully consider the
how the attributes of each of thousands of stocks satisfy her own preferences and beliefs.
Odean (1999) and Barber and Odean (2008) propose that most individual investors solve
this daunting search problem by choosing from the small subset of stocks to which their
attention has been directed. If the stocks to which an investor’s attention is directed do
not include the investor’s best choices or the choices the investor would have made had
she considered the full set of options, attention may greatly influence stock purchase
decisions.
People tend to allocate more attention to more salient choices, that is, to choices
that differ most noticeably on observed attributes. We propose, however, that which
stocks individual investors are aware of and which they ignore is determined primarily by
media coverage and location, not by salience. Salience influences how an investor’s
attention is allocated among the choices presented. But media coverage and the location
of that coverage determine what those choices are.
For example, an investor may choose to read the Wall Street Journal, but the
Journal’s editors decide which companies are covered in the Journal and whether those
companies appear on the front or back pages. For many, if not most, investors, this
filtering by information intermediaries matters more than the salience of securities that
pass through the filter. Furthermore, information intermediaries may create salience
where it did not previously exist.
One illustration of creating salience is the Wall Street Journal’s “Dartboard”
column, published from October 1988 through April 2002. Every month, four investment
professionals each recommended one stock pick. These picks were pitted against four
stocks chosen by darts thrown at stock tables on the Wall Street Journal’s walls. <
https://www.wsj.com/articles/SB10190809017144480> Barber and Loeffler (1993)
investigate the performances of stocks covered by the “Dartboard” column from Wall
Street Journal from October 1988 through October 1990. They find that the stocks
3
recommended by the “Pros” experienced average positive abnormal returns of 4 percent
and double their average trading volume on the two days following the publication of the
recommendations. The recommended stocks with the highest trading volumes
experienced the highest abnormal returns followed by significant reversals from days 2 to
25.
What brought these stocks to investors’ attention was not their inherently salient
attributes. Of course, the investment professionals may have chosen to recommend stocks
with a compelling narrative, but in most cases, nothing newsworthy had happened to
recommended stocks since the previous day’s issue of the Journal and these stocks were
not receiving unusual coverage in other news outlets. Thus many of the recommended
stocks probably did not have attributes that would have attracted the attention of an
investor who was scanning information about the entire universe of stocks on his own.
What brought these stocks to investors’ attention was that they were mentioned in the
Journal.
Furthermore, relative to other stocks mentioned in the Journal that day, these
stocks were salient, but, again, not because of their inherent attributes. They were salient
because they were featured in a prominent, popular, narrative-driven column. The
Dartboard column included short bios and a sketch of each expert and as well as a
description of the expert’s rationale for his or her pick. The column was framed as a
contest and readers knew they could look forward to reading in six months1 about how
the experts had fared relative to the darts.
On any day, the thousands of companies listed on US exchanges will not get
equal media coverage. Different financial media may highlight stocks for different
reasons. Some may cover stocks with attributes that investors would, on their own, find
salient such as extreme returns. Some may focus on less salient, yet important,
fundamental information that investors might otherwise miss. And some may run stories
that simply sell newspapers or increase TV viewership. How stories are packaged also
matters. Though Jim Cramer does not mince words when describing the five biggest
winners of 2017 on CNBC’s Mad Money, he leaves number 6 unmentioned. 2 And while
1 Prior to 1990 expert and dart pick returns were reported after one month. 2 https://www.cnbc.com/2018/01/03/cramer-reviews-the-dows-biggest-winners-and-losers-for-2017.html
4
an investor searching all stocks on his own might find the 6th biggest winner nearly as
salient as the 5th, Cramer’s viewers are unlikely to give number 6 any thought.
In this article, we review recent papers on investor attention, focusing primarily,
but not exclusively on individual investors. We discuss how these papers do, or do not,
support our hypothesis that coverage by the media is the primary determinant of investor
attention. Of course, there are other channels through which a stock may attract investors’
attention. For example, an investor may drive by a company’s factory, shop in company’s
store, or purchase a company’s product. Or a friend, co-worker, or brother-in-law may
recommend a stock. However, while local factories and stores may contribute to
individual investors’ home bias in ownership (e.g., Huberman, 2001; Grinblatt and
Keloharju, 2001; Ivkovich and Weisbenner, 2005; Seasholes and Zhu, 2010), these
alternative channels are not likely to result in the systematic short-term changes in
investor trading documented in the articles discussed below.
The papers we review examine different measures of investor attention. Some
document increased trading volume in conjunction with attention grabbing events. Some
document short-term price moves or short-term price moves followed by reversals.
Others measure attention more directly by analyzing measures Internet search volume.
Most studies we discuss document some, but not all, of these effects. We analyze 1983-
2000 ISSM/TAQ and 2007-2017 TAQ data to show that, for several types of previously
studied attention-attracting events, individual investors are on the buy side of the market.
I. Salience Attention is a selective process in which an “organism appears to control the
choice of stimuli that will be allowed, in turn, to control its behavior.” (Kahneman,
1973) Researchers have studied extensively the features of stimuli that attract our
attention and circumstances under which some stimuli are or are not favored. Though our
capacity for attention varies with arousal and effort, it is limited. (Kahneman, 1973).
Stimuli compete for limited resources (Triesman 1960) and that limit can prevent us from
attending to all available stimuli or even all stimuli relevant to a task. Attention is
directed towards salient stimuli that stand out from the background on dimensions such
5
as movement, color, brightness, or, on a higher cognitive level, unexpectedness. Salient
stimuli have proportionately more influence on behavior (Shinoda, Hayhoe, and
Shrivastava 2001).
Odean (1998) argues that investors “overweight salient, anecdotal, and extreme
information,” relative to abstract, statistical, and base-rate information, and that
overweighting probabilities associated with salient information—such as recent extreme
returns--and underweighting abstract information—such as earnings announcements--
leads to systematic market over- and under-reactions.
Bordalo, Gennaioli, and Shleifer (2013) develop a salience based model of lottery
selection in which the payoffs of a lottery that differ most noticeably from other the
payoffs of other available lotteries are salient by contrast. The probabilities of salient
lotteries are overweighted. Bordalo, Gennaioli, and Shliefer 2013, argue that salience
theory can explain investor preference for stocks with positive skewness, and for growth
stocks as well as the equity premium and countercyclical variation in market returns.
Our contention is that salience affects investor attention but only for the choices
presented to investors and that media coverage and location are the main determinants of
the choices investors see.
II. Data and Methods
II.A. Identifying individual investor and institutional investor
trades
We identify individual and institutional trades using tick-by-tick transaction level
data for US stock markets from the Trade and Quotes (TAQ) and Institute for the Study
of Security Markets (ISSM) transaction data over the periods 1983–2000 and 2007-2017.
For the 1983-2000 period we rely on algorithms developed by Lee and Ready
(1991) to sign trades as buyer or seller initiated. Following Lee and Radhakrishna (2000),
we define trades of less than $5,000 as individual 3 and greater than $50,000 as
institutional; Lee and Radhakrishna (2000) find that in the 1990-91 TORQ database, 39%
3 All of our results are qualitatively unchanged if we define small trades as less than $10,000 (1991 dollars).
6
of individual investor trades are for less than $5,000 and only 10% of trades less than
$5,000 are institutional, while 35% of institutional trades are for more than $50,000 and
only 2% of trades greater than $50,000 are from individuals. To account for changes in
purchasing power over time, our trade size cut-offs are based on 1991 real dollars and
adjusted annually using the Consumer Price Index (CPI). See Barber, Odean, and Zhu
(2009) and Hvidkjaer (2008) for details). Our analysis ends in 2000 because the use of
computerized trading algorithms to break up institutional trades renders small trade size a
less reliable proxy for trades of individual investors after 2000.
For the period 2007-2017, we identify individual investor purchases and sales
from resulting from marketable orders using methodology developed and described in
Boehmer, Jones, and Zhang (2017).
The 1983-2000 and 2007-2017 methodologies differ in two important respects.
First, for the 1983-2000 period we identify trades of individual investors as buyer or
seller initiated. From 2007 to 2017, we identify trades of individual investors as
purchases or sales. In aggregate, the dollar value of all purchases and all sales of a stock
on a day must be equal. However, there is no such adding up constraint for the dollar
value of buyer and seller initiated trades; a greater dollar value of executed trades can
result from either buyer initiated trades or seller initiated trades. Second, for the 1983-
2000 period, we identify large trades that were most likely initiated by institutional
investors. We do not attempt to identify institutional trades for the 2007-2017 period.
Though Boehmer, Jones, and Zhang (2017) write that they believe their approach picks
up a majority of overall retail trading, the complement to these trades would include
many retail trades.
II.B. Calculating order imbalances
To measure buy-sell imbalances, we follow Barber and Odean (2008) by forming daily
(or weekly) portfolios of stocks based on sorting criteria such at abnormal trading volume
or days in event time. For example, to calculate the buy-sell imbalance for small trades,
for each stock on each trading day we calculate the stock’s abnormal trading volume as
the ratio of the stock’s trading volume that day to its average trading volume over the
previous one year (i.e., 252 trading days). We sort stocks into vigntiles on the basis of
7
that day’s abnormal trading volume. Then we sum the number of small buys (B) and
small sells (S) of stocks in each volume partition on day t and calculate buy-sell
imbalance for purchases and sales executed that day as:
1 1
1 1
pt pt
pt pt
n n
it iti i
pt n n
it iti i
NB NSBSI
NB NS
= =
= =
−=
+
∑ ∑
∑ ∑ (1)
where npt is the number of stocks in partition p on day t, itNB is the number of
purchases of stock i on day t, and itNS is the number of sales of stock i on day t. We
calculate the time series mean of the daily buy-sell imbalance (BSIpt) for the days that we
have trading data for each investor type. Note that our measure of buy-sell imbalance
considers only executed trades; limit orders are counted if and when they execute. If
there are fewer than ten trades in a partition on a particular day, that day is excluded from
the time series average for that partition. We separately calculate daily buy-sell
imbalance for large trades (1983-2000) and retail trades (2007-2017). An analogous
procedure is followed to calculate weekly buy-sell imbalances based on contemporaneous
weekly abnormal volume sorts.
For daily return sorts, each day (t-1), we sort all stocks for which returns are
reported in the CRSP NYSE/AMEX/NASDAQ daily returns file into vingtiles based on
the one day return. We calculate the time series mean of day t daily buy-sell imbalance
(BSIpt) for the days that we have trading data. An analogous procedure is followed to
calculate weekly buy-sell imbalances based on the previous week’s returns.
We also sort stocks daily based on abnormal news intensity. To be included in
this analysis, a stock must have at least 10 stocks ticker appearances in the daily news
feed from the Dow Jones News Service during the formation period of 207 to 21 days
prior to the sorting day. Included stocks without news in the current day are assigned to
bin 0. Stocks with news on the current day, t=0, are sorted into quartiles based on the
abnormal news measure of (the number of news stories in current day) / (average daily
number of news stories from day t = -270 through t = -21. We calculate daily (BSIpt) and
its time series average.
8
We calculate buy-sell imbalances for event day partitions of events analyzed in
some of the papers discussed below. Each event e is identified by a stock-date pair: stock
k(e) , and date t(e) (e.g., stock k is recommended in the WSJ’s dartboard column on
date t). We calculate the small trade buy-sell imbalance for event e on event day t as
BSIet =
NBk (e)t (e) − NSk (e)t (e)
NBk (e)t (e) + NSk (e)t (e)
(2)
where NBk (e)t (e) is the number of small buys and
NSk (e)t (e) the number of small sells of
stock k(e) on event date t(e) . We then calculate the average buy-sell imbalance for all
events on day t. If there are fewer than ten trades in a stock on event day t, that stock is
excluded from the buy-sell imbalance average for that day. We use the same procedure to
calculate small trade buy-sell imbalance for event days t-10 through t+10. We separately
calculate event time buy-sell imbalances for large trades (1983-2000) and retail trades
(2007-2017).
From 1983-2000, there are 471 million small buys, 455 million small sells, 190
million large buys, and 172 million large sells. The average small buy size is $3,179, the
average small sell, $3,106, the average large buy, $227,786, and average large sell,
$229,749. From 2007-2017, there are 1.60 billion retail buys and 1.52 billion retail sells.
The average retail buy size is for $15,203 and the average retail sell for $15,626.4 Each
trading day (1983-2000), we calculate the number of small purchases, small sales, and the
buy-sell imbalance (i.e., equation 2) for small trades that day. We do the same for large
trades from (1983-2000) and for retail trades (2007-2017). Table 1 provides means,
medians and other descriptive statistics for these calculations. The time series average
buy-sell imbalance for small trades is 2.89% and for retail trades 2.35%. Large trades
lean more heavily towards buying with a time series average buy-sell imbalance of
5.12%.
III. The asymmetrical effect of attention on buying and
selling Barber and Odean (2008) propose that attention increases the buying and selling
4 The average trade size calculations for small, large, and retail trades are not adjusted for inflation.
9
of individual investors asymmetrically and that attention driven buying by individuals
creates price pressure that temporarily increases prices and is followed by lower
subsequent returns.
When buying stocks, investors face a huge search problem. There are thousands
from which to choose. Investors who do not systematically search for stocks are likely to
make purchases from the subset of stocks that catch their attention. Even news that most
investors consider negative will result in more purchase volume if a large number of
investors hear the news and some take a contrary view.
Selling is different. Individual investors usually own only a few individual stocks
and most do not sell short (Barber and Odean, 2000). While negative news could prompt
an investor to sell, only the small subset of individual investors who already own the
stock are likely to sell; furthermore, these investors probably think about, and consider
selling, stocks they already own even on days when those stock aren’t in the media.
Thus news or other media attention—both good and bad—create a short-term
imbalance in the buying and selling of individual investors that tends to favor buying.
Attention also matters for institutional investors, but less so than for individuals.
Institutions have more attention. They are more likely to work in teams, work long hours,
and to use computers when searching for stocks to buy or to sell. Institutions typically
own more securities than individuals and many institutions do short sell. Thus attention
does not affect the buying and selling of institutions in the same asymmetrical way that it
affects individual investors.
Barber and Odean (2008) test the hypothesis that individual investors are net
buyers of stocks on days when their attention is drawn to a stock using three proxies for
investor attention and three databases of individual investor trading records. Their proxies
for attention are: high abnormal daily trading volume— if an unusual number of investors
trade a stock, it is nearly tautological that an unusual number are paying attention to that
stock; extreme positive or negative price moves the previous day; and whether a stock is
mentioned in the Dow Jones News Service that day. Their individual investor databases
are: 78,000 investors at a large discount brokerage (Jan 1991- Dec 1996) (the LDB
dataset), 14,667 investors at a small discount brokerage (Jan 1996 – Jun 1999), and
665,533 investors at a large retail brokerage (Jan 1997 – Jun 1999). They find that for
10
individual investors the average buy-sell imbalance (as defined in Equation 1) is greater
on high attention days. 5 An analysis of a smaller set of professional money managers
does not yield similar patterns.
We extend this analysis by calculating daily buy-sell imbalances associated with
abnormal daily volume, prior day returns, and abnormal daily news coverage for small
and for large trades for the 1983-2000 ISSM/TAQ data and retail trades for the 2007-
2017 TAQ data. We also calculate weekly buy-sell imbalances for stocks sorted on the
same week’s abnormal volume, the previous week’s return, and same week’s abnormal
news coverage.
Figure 1 graphs buy-sell imbalances for vingtiles sorted on abnormal daily and
weekly abnormal trading volume. Consistent with Barber and Odean’s (2008) theoretical
model and brokerage data results, small trade and the retail trade imbalances are
increasing monotonically and, significantly, in the abnormal volume sorts. Individual
investors place progressively more buy orders relative to sells when abnormal trading is
higher. Note that they are not simply trading more; they are buying more. The increases
are remarkable smooth. The range of the imbalances is tighter for the retail trades,
possibly because these include larger retail trades or because of changes in the trading of
retail investors over time. The pattern for large trades suggests that the trading of
institutional investors is less affected by attention and institutions buy stocks with the
highest and lowest abnormal trading volume less aggressively. Sorting stocks on weekly,
rather than daily, abnormal volume yields very similar results.
Figure 2 graphs daily (and weekly) buy-sell imbalances for vingtiles sorted on the
previous days (or previous week’s) return. The results are again consistent with Barber
and Odean’s (2008) theoretical prediction and brokerage data results. Individual investor
buy-sell imbalances are U shaped when sorted on previous period return; that is,
individual investors make proportionately more purchases than sales of stocks with
extreme recent returns, whether these returns are positive or negative. In contrast,
institutions make proportionately fewer purchases of stocks with extreme recent returns.
The buying activity of individual investors, but not institutions, appears to be driven by 5 Barber and Odean (2011) report empirical evidence that stocks tend to underperform the month following high attention days.
11
attention. As in Figure 1, retail trades for the period 2007-2017 display a similar pattern
to small trades for 1983-2000, however the range of imbalances is tighter.
The abnormal trading volume results and the previous period return results
confirm that attention leads individual investors to be disproportionately on the buy side
of the market. However, abnormal trading volume is a measure of investor attention but
does not tell us why investors are paying attention. And investors may pay attention to
stocks with extreme recent price moves because these stocks are mentioned in the media
or because such moves are salient. In Figure 3 we look at the influence of news on
investors’ attention and buy-sell imbalances.
Figure 3 graphs daily and weekly buy-sell small and large trade imbalances for
abnormal news sorts. On both a daily and weekly basis, the trades of individuals are least
likely to be purchases for stocks not mentioned in the news and most likely to be
purchases for stocks with high abnormal news coverage. In contrast, institutional trades
are least likely to be purchases when abnormal news is high. Individuals, but not
institutions, appear to be buying stocks to which their attention is directed by the media.
III.A. Location
One striking example of the importance of where and how an event is reported in
the media is Huberman and Regev’s (2001) study of EntreMed. On Sunday, May 3, 1998,
the New York Times ran a front-page article about a breakthrough in cancer research and
mentioned Entre Med (ENMD) a company with licensing rights to the breakthrough. The
share price of the stock soared from the previous Friday’s close of 12 to 85 Monday
morning and closed Monday at 52. However, the development of the potential cancer
cure had been reported in an article in Nature five months ago on November 27, 1997.
Furthermore, the Nature article had at that time been covered in the New York Times as
well as CNN and CNBC and was the subject of an Entre Med news release. The price of
ENMD moved from a closing price of 11.75 on November 26th to a closing price of
15.25 on November 28th. Thus the market reaction to a prominently placed story with no
new information was much greater than the earlier reaction to the actual new news.
Location dominated content. But not similarly for all investors.
In Figure 4, we graph the cumulative number of small purchases minus small
12
sales of EntreMed from October 1997 through December 1998 (the same period graphed
in Figure 1 of Huberman and Regev (2001)). We do the same for large trades. Individuals
increase buying somewhat in the week following the publication of the Nature article on
November 26, 1997. On November 26, there were 15 more buyer initiated small trades
than seller initiated small trades; and over the next two weeks 261 more buyer initiated
small trades than seller initiated small trades. On November 26 there were 0 more buyer
initiated large trades than seller initiated large trades and over the next two weeks 1 more
buyer initiated large trades than seller initiated large trades. However the reaction of
individuals to the Nature article was tiny compared to the huge spike small trade net buys
on May 4th, the day after the New York Times article appeared. On May 4th, there were
2,898 more buyer initiated small trades than seller initiated small trades; and over the
next two weeks an additional 1,729 more buyer initiated small trades than seller initiated
small trades. In contrast, on May 4th there were only 62 more buyer initiated large trades
than seller initiated large trades and over the next two weeks 150 fewer buyer initiated
large trades than seller initiated large trades. The New York Times published; individuals
bought.
While Huberman and Regev (2001) demonstrate the importance of where and
how and event is reported in the media, Engelberg and Parsons’ (2011) study of local
newspaper coverage of earnings reports focuses on geographic cross-sectional differences
in news coverage of the same event. To separate the effect of media coverage from
investor’s reaction to the underlying event, Engelberg and Parsons (2011) identify 19
local newspapers serving major U.S. cities and examine the trading of individual
investors living in or near those cities. The study analyzes trading in the LDB dataset
around S&P 500 Index firm earnings announcements.
The authors first determine whether each local newspaper reported an earnings
announcement within two days of the announcement. They then analyze whether local
news coverage explains differences in local investor trading in earnings announcement
firms. After controlling for a host of explanatory variables—including firm-city, firm-
earnings date, and city-earnings date fixed effects, they find that local newspaper
coverage of a firm’s earnings announcement increases local trading in the firm’s stock
even for non-local stocks. There is, of course, some possibility that local newspapers
13
simply report earnings announcements that of more interest to their readership. To
control for this possibility, Engelberg and Parsons (2011) run a second analysis of local
investor trading in response to earnings announcements on days on which a local
newspaper’s delivery was impeded by hail or blizzards. While neither hail nor blizzards
significantly suppress local trading in general, they sever the relationship between local
newspaper coverage of an earnings announcement and local investor trading. If the local
newspaper covers the announcement but the newspapers are not delivered, the effect on
local individual investor trading is similar to what happens when the local paper does not
report an earnings announcement.
Peress (2014) reports that national newspaper strikes lead to lower trading
volume, volatility, and the dispersion of stock returns. He finds that small-cap stocks are
most affected by strikes. Peress (2014) also documents a local drop in individual investor
trading (LDB data) on days of and in regions of local newspaper strikes in the US..
Fedyk (2018) demonstrates the importance of news location for institutional
investors in a context in which the limits of institutional investors’ attention are taxed.
Bloomberg’s editors classify news articles as “primary important” (PI) “secondary
important” (SI) or “all other.” Primary important articles get “pinned” to the top of
Bloomberg terminal screens. Below this, SI and other articles scroll down the screen as
they are replaced by new articles. At most three articles are pinned to the top of the
screen. Articles remain there until they are either replaced by a new PI article or “a
predefined amount of time (on the order of hours) elapses.” If a PI article hits its time
limit without being replaced by another PI article, the next SI article to be published is
pinned to the top of the terminal screen. SI articles are selected to be pinned solely
because they were in the right place at the right time; they are not selected based on
importance relative to other SI articles.
Fedyk (2018) measures the market reactions to 1,274 SI articles that are pinned to
the top of the screen and 8,233 SI articles that are not pinned (from March 2014 to
December 2015). In the ten minutes following publication, stocks mention in pinned SI
articles experience 280% higher trading volume and 180% large price changes than
stocks mentioned in unpinned SI articles. Stocks in pinned articles have strong return
drift for the next 30 to 45 minutes but no subsequent drift. Stocks in unpinned articles
14
experience much smaller initial market reactions and drift that continues for 10 to 15 days
at which point the average returns to positive and negative pinned and unpinned news are
virtually identical.
Not only are pinned SI articles prominently located, but, on average, they remain
on the screen much longer than unpinned articles. It is possible that some investors who
look at the Bloomberg screen and see both a pinned SI article and unpinned SI article are
attracted by the salience of the pinned article. However, it is almost certain, that more
investors see the pinned SI articles than the unpinned SI articles. Even for institutional
investors, the location of media coverage matters.
III.B. Media Recommendations
Griffin and Tversky (1992) write that “people focus on the strength or
extremeness of the available evidence (e.g., the warmth of a letter or the size of an effect)
with insufficient regard for its weight or credence (e.g., the credibility of the writer or the
size of a sample).” Expert stock recommendations reported in the media are a case in
point. These recommendations rarely convey new information nor accurately predict long
run performance. But they do garner attention.
As discussed above, Barber and Loeffler (1993) find that stocks recommended
from 1988 to 2000 by experts in the Wall Street Journal’s “dartboard column”
experienced positive abnormal returns of 4 percent during the two days of trading
beginning the publication date of the column; “in contrast, the Dartboard Stocks
experience no significant abnormal returns over the same period.” The abnormal returns
are followed by partial reversals over the next 25 trading days. Wright (1994) reports a
similar finding for the first 20 publications of the dartboard column with a 4.59 percent
abnormal return in the two-day publication window and a complete reversal after 36
trading days. Metcalf and Malkiel (1994) examine the dartboard column expert
recommendations for 1990 through 1992. They find a one-day announcement effect of
3%, which they attribute to publicity, but no significant evidence of long-term
outperformance by the experts. Liang (1999) examines the performance of dartboard
expert recommendations and dart picks for 1990 through November 1994. For expert
recommendations, he documents a 3.5% abnormal return during the 2-day announcement
15
period; “for the dartboard stocks there is no such price reaction following the
announcement.” Liang estimates that during the 6-month contest and after adjusting for
risk, investors following the experts’ recommendations lose 3.8%. Both Barber and
Loeffler (1993) and Liang (1999) report a sharp increase in trading volume during the
announcement window for the expert recommended stocks but not for the dart picks.
In Figure 5, we graph the buy-sell imbalance of individuals and institutions prior
to following the announcement of expert dart board column recommendations and dart
picks. There is a sharp increase in the buy-sell imbalance of individuals the day of expert
recommendations, decaying over the next week, but no such increase in response to dart
picks. Institutional buy-sell imbalance does not react to either the darts or the experts.
Thus individual investors appear to be buying stocks recommended in the press although
these recommendations do not reveal novel news nor lead to positive abnormal long run
returns. Investors do not, however, buy the stocks chosen buy darts even though dart
picks are reported in the same column; buying a stock picked by a dart is difficult to
reconcile with a self-image as a serious investor.
Several studies look into the recommendations given on the Wall $treet Week, an
investment TV show originally hosted by Louis Rukeyser aired each Friday evening on
PBS in the United States from 1970 to 2002. Recommendations, both positive and
negative, were made by regular panelists and by weekly guests, typically successful
analysts or money managers.
Pari (1987) documents that the stocks recommended on Wall $treet Week during
1983-1984 experienced positive abnormal returns the week after being recommended, but
underperformed the market by an average of 1% within two weeks.
Beltz and Jennings (1997) analyze how 800 recommendations made on Wall
$treet Week during 1990-1992 influence returns and the volume of buyer initiated trades
by all market participants.6 Positive recommendations have positive abnormal returns the
week following the broadcast, with strong positive recommendations averaging a
cumulative abnormal return (CAR) of 3.93%. These positive returns are followed by
reversals with an average CAR of 0% (-1.18%) three (six) months after the broadcast.
6 Beltz and Jennings (1997) estimate buyer initiated trading volume from the Institute for the Study of Securities Markets (ISSM) quote and trade data.
16
Positive abnormal buying volume is also observed the week following the broadcast.
Negative recommendations have -0.62% abnormal return the trading after the
recommendation and a CAR of -3.25% after six months. No abnormal selling volume is
observed.
Ferreira and Smith (2003) analyze over 350 recommendations on Wall $treet
Week for the period 27 December 27, 1996 through December 26, 1997) by matching
companies in the sample with control groups of companies with similar size, book-to-
market, and industry characteristics. They document an abnormal return of 0.651% the
trading day after the broadcast followed by a reversal. However, they estimate that on
average recommended stocks outperform stocks with similar characteristics over the two
years following the recommendations.
Mad Money with Jim Cramer is a stock market oriented television program
broadcast on CNBC Monday through Friday evenings since 2005. Engelberg, Sasseville
and Williams (2012) examine the market reaction to 826 first-time buy recommendations
on Mad Money from July 28, 2005 to February 6, 2009. They construct portfolios of
recommended stocks 2 hours before the show airs at 6PM (EST) to capture the price
change that follows the recommendation. These portfolios experience an average
overnight return of 2.4% followed by a return reversal. This return reversal effect is
strongest for 1) stocks otherwise not in the news (i.e., not appearing in a non-Cramer
related article in Factiva during a three-day window surrounding the recommendation
date); 2) small illiquid stocks with high idiosyncratic volatility and high Specialness (i.e,
stocks that are difficult to arbitrage); and 3) stocks recommended on evenings when the
show’s Nielsen Ratings are higher. The correlation between Nielsen Ratings and the
overnight return of recommended stocks is strongest when viewership is greater in high-
income area. The authors also examine the market reaction to Cramer’s negative
recommendations. Consistent with the hypothesis that the buying and selling of
individual investors respond asymmetrically to attention, stocks with negative
recommendations experience a small negative overnight average return of -0.29%
without a subsequent reversal.
Neumann and Kenny (2007) examine short-term market responses to 106 buy
recommendations and 46 sell recommendations made on Mad Money with Jim Cramer
17
between July 26, 2005 and September 16, 2005. They find an average abnormal return of
1.06% from the close of trading the day of a buy recommendation (t = 0) to the close on
the following trading day (t = 1) with most of this return between the close on day t and
the open on day 1. They also find statistically weak evidence of a subsequent reversal.
Sell recommendations are followed by negative one day return of -0.20% and no reversal.
Lim and Rosario (2010) analyze 10,589 recommendations made on Mad Money
with Jim Cramer between June 28, 2005 and December 22, 2006. The sort
recommendations as buy or sell, by the market cap of the recommended firm, and as
“caller” or “non-caller” recommendations. Caller recommendations are those made by
Cramer in response to viewer phone calls. Lim and Rosario (2010) find that both buy and
sell recommendations have positive cumulative excess returns in the month leading up to
the recommendation. The one-day positive market reaction to buy recommendations and
negative reaction to sell recommendations is strongest for small capitalization stocks. In
contrast to Engelberg, Sasseville and Williams (2012) and Neumann and Kenny (2007),
Lim and Rosario (2010) find weak evidence that Cramer’s non-caller recommendations,
particularly for smalls stocks, are positively correlated with returns over the next six
months.
Karniouchina, Moore, and Cooney (2009) analyze 7,160 buy recommendations
between made on Mad Money with Jim Cramer November 1, 2005 and July 31, 2007.
Their study focuses on attributes of these recommendations known to influence the
efficacy of paid advertisements, including primacy-recency effects, clutter and
competition, and message length. Like other studies, they document positive pre-
recommendation returns, strong one-day post recommendation returns, and subsequent
reversals. They find greater one-day market reactions to stocks recommended at the
beginning (and to a lesser extent at the end) of a program segment and to stocks receiving
more detailed recommendations. They do not find that stocks react more strongly when a
program has fewer total recommendations.
Keasler and McNeil (2010) study 7,807 recommendations made on Mad
Money with Jim Cramer between December 1, 2005 through December 31, 2006. They
document a strong increase in trading volume following non-caller recommendations; the
increase trading volume is particularly pronounced for small capitalization stocks. One
18
day market reactions to non-caller buy recommendations are positive especially so for
small capitalization stocks and are followed by reversals over the next 25 trading days.
One day market reactions to non-caller sell recommendations are less pronounced than to
non-caller buy recommendations; they are negative for mid and small capitalization
stocks but virtually 0 for large stocks; reversals are also less pronounced for non-caller
sell recommendations than buy recommendations. Keasler and McNeil (2010) conclude
that the initial market reactions to recommendations and subsequent reversals are
consistent with the price pressure hypothesis.
III.C. Advertising
Though most advertising promotes sales of specific products, advertising can also
direct the attention of an investor to a company. More attention increases the likelihood
that more investors will buy a stock (while affecting selling less). Grullon, Kanatas and
Weston (2004) document that the magnitude of a firm’s annual advertising expenditures
correlates with greater breath of stock ownership as well as greater trading liquidity.
Analyzing consumer survey data from the Landor Image Power, Frieder and
Subrahmanyam (2005) find that the proportion of a firm’s shareholders who are
individual investors is greater for firms with high brand familiarity and high regard for
brand quality.
Lou (2014) analyzes the effect of changes in a firm’s annual advertising
expenditures on the firm’s stock performance and on the buy-sell imbalance of individual
investors. He finds that firms in the top decile of year-to-year changes in advertising
spending outperform those in the bottom decile by 12.58 percent in the ranking year and
underperform by 6.96 percent and 9.84 percent in the following two years respectively.
Lou calculates monthly small trade buy-sell imbalances for the 1983-2000 ISSM/TAQ
data signing trades with the Lee Ready algorithm. He finds that retail investors place
proportionately more buy orders for firms that increase their advertising. Lou also
presents evidence that managers are aware of exploiting the link between changes in
advertising and stock returns. He finds that a firm’s advertising tends to increase in
anticipation of insider sales, seasoned equity offerings, and stock-financed acquisitions.
19
Focke, Ruenzi, and Ungeheuer (2014) focus on firms’ intraday advertising
expenditures for TV, daily expenditures for newspaper and magazines, and monthly
Internet, radio, and billboard expenditures compiled by Kantar Media from 1995 through
2012. They measure investor attention using the number of views of a company’s
Wikipedia page. They find a strong positive correlation between a firm’s abnormal
advertising spending and both contemporaneous and subsequent Wiki page views.
However, they do not find short-term stock price reactions to abnormal advertising.
While most advertising is designed to directly influence consumers, Focke,
Niessen-Ruenzi, and Ruenzi (2016) document a second channel by which advertising
may influence investor behavior. Firms that spend more on advertising in newspapers
receive more positive coverage and less negative coverage by those same newspapers.
Thus, not only is advertising likely to direct investors’ attention, but it may also lead to
biased reporting that sways investors’ beliefs.
Fehle, Tsyplakov and Zdorovtsov (2005) look stock trading volume and returns
for firms that run commercials during 19 Super Bowls. The classify firms as recognizable
or unrecognizable based on whether the firm running the commercial could be identified
by watching the commercial. They find that recognizable firms, but not unrecognizable
firms, experience positive abnormal returns in a three-day window following the Super
Bowl. For the period 1997-2001, they use small trades in the TAQ data signed by the Lee
Ready algorithm to identify individual investor trades, they find that for recognizable
firms, but not unrecognizable firms, the buy-sell imbalance of individual investors
increases in the three days following the Super Bowl.
Mayer (2016) examines returns and trading in the stock of firms sponsoring
NCAA football bowl. He finds a significant positive abnormal return the trading day (and
week) following the game. Volume increases as does the buy-sell imbalance of individual
investors (estimated using signed small trades in the TAQ data from 1993-2014). The
magnitude of market reaction and volume are positively correlated with the number of
households watching the football game. Mayer also finds a significant increase in Google
search volume (SVI) for game sponsors in the week following the game.
20
III.D. Extraordinary events
Seasholes and Wu (2007) analyze individual investor account level trading on the
Shanghai Stock Exchange, the day of and the day after stocks hit a daily upper price
limit. They find that, on average, individual investors’ buy-sell imbalance (i.e., the value
of individual investor purchases minus sales divided by the value of individual investor
purchases plus sales), is significantly negative the day upper price limits are hit and
significantly positive the following day. Both effects are stronger when fewer upper price
limit events occurred on the same day. Seasholes and Wu (2007) also find that more
investors buy a stock for the first time the day after an upper price limit was reached than
on other days.
Peng and Xiong (2006) develop a theoretical model in which limited attention
causes investors to focus more on information about categories such as the market as a
whole or an industry than on firm specific information. Li and Yu (2005) look at investor
behavior in response to high levels of the Dow Jones Industrial Average Index—a
prominent market statistic. They find that the ratio of the current Dow index to the Dow
52-week high (nearness to 52-week high) positively predicts market returns7 while the
ratio of the current Dow index to the historical high (nearness to historical high)
negatively predicts market returns.
Yuan (2015) examines the response of individual investors to historical highs of
the Dow Jones Industrial Average Index and to front-page news about the market in the
New York Times and the Los Angeles Times. Analyzing the ISSM/TAQ data from 1983-
1999, Yuan finds that for signed small trades (i.e., less than $10,000) seller initiated
orders increase relative to buyer-initiated orders in response to these events. The
imbalance shifts in the opposite direction for large trades (i.e., greater than $50,000).
Yuan also finds that individual investors in the LDB dataset are more likely to sell stocks
when the Dow hits a record high. Furthermore, he sorts stocks on daily abnormal volume
(as in Barber and Odean 2008) and finds that the difference in the buy-sell imbalance for
high and low abnormal volume stocks is greater for days on which the Dow hits a record
high or there is front page market news. Finally, Yuan finds that aggregate mutual fund
7 George and Hwang (2004) find a similar effect for individual stocks.
21
flows (1998-2005) respond negatively to highs in the Dow and to front-page market
news.
Kaniel and Parham (2016) find that quarterly mutual fund flows into funds
mentioned in the Wall Street Journal “Category Kings” top 10 ranking lists are
substantially higher than flows into funds that just miss making the list. The last fund
mentioned in the list experiences nearly a 1/3rd average increase in flows relative to the
next highest ranked fund. Thus media coverage versus non-coverage affects investor
behavior far more than meaningful differences in fund characteristics. Kaniel and Parham
(2016) also show that there is no such increase in flows for funds making similar lists
published less prominently and with less catchy titles in the WSJ. This is consistent with
the hypothesis that the media influence on investor attention depends on coverage and
location as well as salience added by the media.
III.E. Online search as a measure of attention
We contend that the stocks to which investors pay attention is determined
primarily by media coverage and location. Investors choose which newspaper they read,
TV program they watch, or website they visit, but their attention is then directed to one
stock or another by the producers of that content. However, some investors undoubtedly
get their trading ideas from sources other than the media and many investors whose
choices are inspired by media will seek more information before trading.
Da, Engelberg, and Gao (2011) introduce Internet search queries as a measure of
investor attention. Using Google’s weekly Search Volume Index (SVI) they define a
stock’s abnormal Search Volume Index (ASVI) as the log of SVI for a stock’s ticker
during the week minus the log of the median SVI for that ticker during the previous eight
weeks. Consistent with Barber and Odean’s (2008) price pressure hypothesis, Da,
Engelberg, and Gao (2001) find that increases in ASVI predict higher stock prices in the
next two weeks followed by reversals in the next year and that these predictions are
strongest for stocks traded more by individual investors. ASVI also positively correlates
with first-day IPO returns and long-run IPO underperformance.
We examine the relationship between Da, Engelberg, and Gao’s (2011) ASVI
measure and retail investor buy-sell imbalance using 2007-2017 TAQ data. We sort
22
stocks on into vingtiles based on weekly ASVI and calculate average retail buy-sell
imbalance in the same week for stocks in each vingtile. Average buy-sell imbalances with
95% confidence intervals are graphed in Figure 7 where retail buy-sell imbalance
increases nearly monotonically in ASVI. Once again, individual investors are on the buy-
side of the market for stocks that catch their attention.
Joseph, Wintoki, and Zhang (2011) also look at the relationships between Google
search volume, trading volume, and returns. Their measure of search intensity is
Google’s weekly normalized and scaled search volume intensity for an S&P 500 stock’s
ticker. They find that higher search volume in a week predicts higher abnormal trading
volume and abnormal returns the following week and that the sensitivity of returns to
search volume is greater for stocks that are more difficult to arbitrage.
IV. Inattention We review the literature that studies events for which the attention of investors—
primarily individuals—is attracted to a stock. There is a separate literature on investor
inattention. While attracted attention often results in short-term overreactions followed by
reversals, inattention leads to underreactions and continued drift.
Influential papers in the investor inattention literature include Hirshleifer and
Teoh (2003), Hirsheilfer, Lim, and Teoh (2011), Della Vigna and Pollet (2009), and
Hirshleifer, Lim, and Teoh (2009). Hirshleifer and Teoh (2003) and Hirsheilfer, Lim, and
Teoh (2011) develop models in which limited investor attention leads to initial under-
reaction to earnings announcements and other accounting data followed by drift. Della
Vigna and Pollet (2009) show that the initial stock reaction to earnings announcements is
smaller and post-earnings announcement drift greater for firms that announce earning on
Fridays—when inattention is more likely—than on other weekdays. Hirshleifer, Lim, and
Teoh (2009) document that the initial reactions are smaller and post-earnings
announcement drift greater for firms announcing earnings on days when many other
firms make earnings announcements.
While inattention is often driven by too much competition for investors’ attention,
it can also result from a lack of insight. Cohen and Frazzini (2008) document that when
firms have strong customer-supplier economic links, the stock of one firm will underreact
23
to negative economic news about the other firm. DellaVigna and Pollet (2007) show that
stock prices underreact to demographically predictable changes in future demand for
products targeted to specific age groups if those changes are 5 to 10 years in the future.
Both of these papers analyze examples for which relevant information was available to
investors, but investors appear to have not fully appreciated its importance. By
highlighting the potential profits from paying closer attention this information, these
papers may lead to a reduction in the inattention that they document.
V. The Media is Changing When Loius Rukeyser first hosted Wall $treet Week November 20, 1970,8 the
Public Broadcasting System had been broadcasting for only six weeks. National
television was dominated by three networks. Basic cable networks did not yet exist.
Investors who wanted to learn about stocks had few choices. They watched Rukeyser on
Friday nights. They read the business section of their local newspaper.
By the early 1980s, 4 million US households were watching Wall $treet Week,9
double the circulation of the Wall Street Journal.10 As shown in Engelberg, Sasseville and
Williams (2012), Engelberg and Parsons (2001), and Peress (2014), the influence of a TV
show or newspaper on market prices depends on how many people watch or read it.
Rukeyser and his guests influenced stock prices because a substantial proportion of active
investors were watching.
Today investors have thousands of sources of information about stocks: 24-hour
cable financial news, online newspapers, searchable commercial websites devoted to the
market, SEC websites, blogs, market oriented social media. Investors can choose what
type of information they see. If investors are fragmented in the media sources to which
they attend, the influence of any one source on aggregate investor behavior is likely to
diminish.
Wall $treet Week broadcast on Friday nights. The stock market opened Monday.
Investors had time to mull over their decisions and, perhaps, do research. Today most
8 https://www.nytimes.com/2006/05/03/business/media/03rukeyser.html 9 https://www.nytimes.com/1990/11/11/business/enduring-not-always-endearing-wall-street-week.html 10 In 1982 the Wall Street Journal had a national circulation of 1,927,963 (’82, Ayer Directory of Publications, 1982, IMS Press)
24
sources of market information are available while markets are open. Investors can react
immediately and may feel an urgency to do so. Thus the market reaction to the media’s
influence on investor attention is likely to be faster than in the past. Moreover, investors
in a hurry to trade may not take the time to verify what they just read and may become
increasingly vulnerable to rumors or mistaken information.
VI. Conclusion We propose that the stocks to which individual investors allocate their attention
are determined primarily by media coverage and location. We review dozens of papers on
investor attention with an eye to examining our hypothesis and find strong support.
Stocks with no media coverage get little attention. For example in Figure 4, stocks
with no news coverage have lower average buy-sell imbalances than stocks with
coverage. And Engelberg and Parsons (2001) find that local trading of S&P 500 stocks
on earnings announcement days is lower when the local newspaper is on strike or unable
to be delivered due to weather conditions. And mutual funds that just miss top ten lists
garner fewer inflows than funds that just make the list (Kaniel and Parham (2016)).
Stocks covered in more prominent locations get more attention. Individuals paid
little attention to a Nature article on EntreMed but a lot of attention to a New York Times
article published five months later with same information. Not only did trading volume
and price spike in response to the New York Times article (Huberman and Regev (2001),
but, as illustrated in Figure 1, individuals were dramatically on the buy side of the market
while institutions were not.
Individual investors tend to be net buyers of stocks that get their attention whether
for good or bad reasons. For example, as we see in Figure 2 and in Barber and Odean
(2008), individual investor buy-sell imbalances increase for stocks with both extreme
positive and extreme negative returns in the previous day or week (while institutional
buy-sell imbalances decrease with extreme returns).
Salience plays a role in investor decisions but often only after media coverage and
location pare down the choice set. Sometimes salience is determined by stock attributes
such as earnings surprises or extreme recent price moves. However, salience is often
created by the media itself such as when the Wall Street Journal’s dartboard column
25
experts tell engaging stories about the firms they recommend or Jim Cramer throws a
yellow penalty flag. In Figure 5 we graph the mean buy-sell imbalance for small and
large trades over a 21 event-day window centered on the publication date of the dartboard
column and find that individuals actively trade and are on the buy side of the market for
stocks recommended by experts in the column but not for stocks picked by the darts.
Institutional investors may also be reading the dartboard column, but ignore both the
experts and the darts when they trade.
While it is may not be surprising that individuals trade in response to expert
recommendations (even recommendations institutions ignore), individuals also trade in
response to media coverage with little informational content. The individual investor buy-
sell imbalance for the stocks of firms running Super Bowl commercial increases for three
days after the Super Bowl, but only if the firms is easily recognized from the commercial
(Fehle, Tsyplakov and Zdorovtsov (2005)).
When individual investors pay attention to a stock, they trade it more and tend to
be buyers. If many individuals buy a stock in response to an attention-grabbing event,
there is a short-term price increase often followed by a reversal, especially so for small
stocks that are difficult to arbitrage. Examples include EntreMed (Huberman and Regev
2001), the WSJ’s Dartboard Column (Barber and Loeffler (1993), Metcalf and Malkiel
(1994), Wright (1994) and Liang (1999)), Wall $treet Week (Pari (1987), Beltz and
Jennings (1997) and Ferreira and Smith (2003)), and Mad Money with Jim Cramer,
(Karniouchina, Moore, and Cooney (2009), Kessler and McNeil (2010), Lim and Rosario
(2010), Engelberg, Sasseville and Williams (2012)).
Individual investors can buy any of thousands of stock, but most only buy and
own a few. Searching those thousands for one’s preferred choices is a daunting task.
Barber and Odean (2008) propose that most investors solve this search problem by
choosing from the much smaller set of stocks that catch their attention. While any given
investor’s attention may be caught by an idiosyncratic event such as a remark made by a
neighbor or a factory passed while driving, the systematic direction of investor attention
is mediated by the media.
26
References Barber, Brad M., and Douglas Loeffler. "The "Dartboard" Column: Second-Hand Information and Price Pressure." The Journal of Financial and Quantitative Analysis 28, no. 2 (1993): 273-84. Barber, Brad M., and Terrance Odean. "Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors." The Journal of Finance 55, no. 2 (2000): 773-806. Barber, Brad M., and Terrance Odean. "All That Glitters: The Effect of Attention and News on the Buying Behavior of Individual and Institutional Investors." The Review of Financial Studies 21, no. 2 (2008): 785-818.
Barber, Brad M., and Terrance Odean, 2011, “All that Glitters: The Effect of Attention on the Buying Behavior of Individual and Institutional Investors,” in The Handbook of New Analytics in Finance eds. Gautam Mitra and Leela Mitra, John Wiley and Sons: West Sussex.
Barber, Brad M., Terrance Odean, and Ning Zhu. "Do Retail Trades Move Markets?" The Review of Financial Studies 22, no. 1 (2009): 151-86. Beltz, Jess, and Robert Jennings. "“Wall street week with Louis Rukheyser” recommendations: Trading activity and performance." Review of Financial Economics 6, no. 1 (1997): 15-27. Boehmer, Ekkehart, Charles Jones, and Xiaoyan Zhang. "Tracking retail investor activity." (2017). Bordalo, Pedro, Nicola Gennaioli, and Andrei Shleifer. "SALIENCE THEORY OF CHOICE UNDER RISK." The Quarterly Journal of Economics 127, no. 3 (2012): 1243-285. Cohen, Lauren, and Andrea Frazzini. "Economic links and predictable returns." The Journal of Finance 63, no. 4 (2008): 1977-2011. Da, Zhi, Joseph Engelberg, and Pengjie Gao, 2011, “In Search of Attention,” Journal of Finance 66, 1461-1499. DellaVigna, Stefano, and Joshua M. Pollet. "Demographics and industry returns." American Economic Review 97, no. 5 (2007): 1667-1702. DellaVigna, Stefano, and Joshua M. Pollet. "Investor inattention and Friday earnings announcements." The Journal of Finance 64, no. 2 (2009): 709-749. Engelberg, Joseph E., and Christopher A. Parsons. "The causal impact of media in financial markets." The Journal of Finance 66, no. 1 (2011): 67-97. Engelberg, Joseph, Caroline Sasseville, and Jared Williams. "Market madness? The case of mad money." Management Science 58, no. 2 (2012): 351-364. Fedyk, Anastassia. Front page news: The effect of news positioning on financial markets. Working paper, 2018. Fehle, Frank, Sergey Tsyplakov, and Vladimir Zdorovtsov. "Can companies influence investor behaviour through advertising? Super bowl commercials and stock returns." European Financial Management 11, no. 5 (2005): 625-647.
27
Ferreira, Eurico J., and Stanley D. Smith. “‘Wall $treet Week’: Information or Entertainment?” Financial Analysts Journal, vol. 59, no. 1, 2003, pp. 45–53. Frieder, Laura, and Avanidhar Subrahmanyam. "Brand perceptions and the market for common stock." Journal of financial and Quantitative Analysis 40, no. 1 (2005): 57-85. Focke, Florens, Alexandra Niessen-Ruenzi, and Stefan Ruenzi. "A friendly turn: Advertising bias in the news media." (2016). Focke, Florens, Stefan Ruenzi, and Michael Ungeheuer. "Advertising, attention, and financial markets." (2016). Griffin, Dale, and Amos Tversky. "The weighing of evidence and the determinants of confidence." Cognitive psychology24, no. 3 (1992): 411-435. Grinblatt, Mark, and Matti Keloharju. "How distance, language, and culture influence stockholdings and trades." The Journal of Finance 56, no. 3 (2001): 1053-1073. Grullon, Gustavo, George Kanatas, and James P. Weston. "Advertising, breadth of ownership, and liquidity." The Review of Financial Studies 17, no. 2 (2004): 439-461. Hirshleifer, David, Sonya Seongyeon Lim, and Siew Hong Teoh. "Driven to distraction: Extraneous events and underreaction to earnings news." The Journal of Finance 64, no. 5 (2009): 2289-2325. Hirshleifer, David, Sonya S. Lim, and Siew Hong Teoh. "Limited investor attention and stock market misreactions to accounting information." The Review of Asset Pricing Studies1, no. 1 (2011): 35-73. Hirshleifer, David, and Siew Hong Teoh. "Limited attention, information disclosure, and financial reporting." Journal of accounting and economics 36, no. 1-3 (2003): 337-386. Huberman, Gur, and Tomer Regev. "Contagious speculation and a cure for cancer: A nonevent that made stock prices soar." The Journal of Finance 56, no. 1 (2001): 387-396. Hvidkjaer, Soeren. "Small trades and the cross-section of stock returns." The Review of Financial Studies 21, no. 3 (2008): 1123-1151. Ivković, Zoran, and Scott Weisbenner. "Local does as local is: Information content of the geography of individual investors' common stock investments." The Journal of Finance 60, no. 1 (2005): 267-306. Joseph, Kissan, M. Babajide Wintoki, and Zelin Zhang. "Forecasting abnormal stock returns and trading volume using investor sentiment: Evidence from online search." International Journal of Forecasting 27, no. 4 (2011): 1116-1127. Kahneman, Daniel. Attention and effort. Vol. 1063. Englewood Cliffs, NJ: Prentice-Hall, 1973. Kaniel, Ron, and Robert Parham. "WSJ Category Kings–The impact of media attention on consumer and mutual fund investment decisions." Journal of Financial Economics 123, no. 2 (2017): 337-356.
28
Karniouchina, Ekaterina V., William L. Moore, and Kevin J. Cooney. "Impact of mad money stock recommendations: merging financial and marketing perspectives." Journal of Marketing 73, no. 6 (2009): 244-266. Keasler, Terrill R., and Chris R. McNeil. "Mad Money stock recommendations: market reaction and performance." Journal of Economics and Finance 34, no. 1 (2010): 1-22. Lee, Charles MC, and Balkrishna Radhakrishna. "Inferring investor behavior: Evidence from TORQ data." Journal of Financial Markets 3, no. 2 (2000): 83-111. Lee, Charles MC, and Mark J. Ready. "Inferring trade direction from intraday data." The Journal of Finance 46, no. 2 (1991): 733-746. Liang, Bing. "Price pressure: Evidence from the “dartboard” column." the Journal of Business 72, no. 1 (1999): 119-134. Li, Jun, and Jianfeng Yu. "Investor attention, psychological anchors, and stock return predictability." Journal of Financial Economics 104, no. 2 (2012): 401-419. Lim, Bryan, and Joao Rosario. "The performance and impact of stock picks mentioned on ‘Mad Money’." Applied Financial Economics 20, no. 14 (2010): 1113-1124. Lou, Dong. "Attracting investor attention through advertising." The Review of Financial Studies 27, no. 6 (2014): 1797-1829. Mayer, Erik T. “Investors Attention and Stock Prices: Evidence From a Natural Experiment.” (2016) Working paper. Metcalf, Gilbert E., and Burton G. Malkiel. "The Wall Street Journal contests: The experts, the darts, and the efficient market hypothesis." Applied Financial Economics 4, no. 5 (1994): 371-374. Neumann, John J., and Peppi M. Kenny. "Does Mad Money make the market go mad?." The Quarterly Review of Economics and Finance 47, no. 5 (2007): 602-615. Odean, Terrance. "Volume, volatility, price, and profit when all traders are above average." The journal of finance 53, no. 6 (1998): 1887-1934. Odean, Terrance. "Do investors trade too much?." American economic review 89, no. 5 (1999): 1279-1298. Pari, Robert A. "Wall Street week recommendations: Yes or no?." The Journal of Portfolio Management 14, no. 1 (1987): 74-76. Peng, Lin, and Wei Xiong. "Investor attention, overconfidence and category learning." Journal of Financial Economics 80, no. 3 (2006): 563-602. Peress, Joel. "The media and the diffusion of information in financial markets: Evidence from newspaper strikes." The Journal of Finance 69, no. 5 (2014): 2007-2043. Seasholes, Mark S., and Guojun Wu. "Predictable behavior, profits, and attention." Journal of Empirical Finance 14, no. 5 (2007): 590-610.
29
Seasholes, Mark S., and Ning Zhu. "Individual investors and local bias." The Journal of Finance 65, no. 5 (2010): 1987-2010. Shinoda, Hiroyuki, Mary M. Hayhoe, and Anurag Shrivastava. "What controls attention in natural environments?." Vision research 41, no. 25-26 (2001): 3535-3545. Treisman, Anne M. "Contextual cues in selective listening." Quarterly Journal of Experimental Psychology 12, no. 4 (1960): 242-248. Wright, David W. "Can prices be trusted? A test of the ability of experts to outperform or influence the market." Journal of Accounting, Auditing & Finance 9, no. 2 (1994): 307-323. Yuan, Yu. "Market-wide attention, trading, and stock returns." Journal of Financial Economics 116, no. 3 (2015): 548-564.
Table 1: Descriptive Statistics
Variable N (Days) Mean Std. Dev. Min p25 p50 p75 MaxPanel A: 1983-2000 ISSM/TAQ
Small TradesNumber of Buys 4687 100.49 173.91 1.32 15.76 33.57 74.28 1038.36Number of Sells 4687 97.08 170.73 1.21 15.94 27.97 67.21 1086.72Buy-Sell Imbalance 4687 2.89% 8.41% -38.94% -1.84% 3.76% 8.50% 41.18%
Large TradesNumber of Buys 4687 40.44 44.00 1.12 12.35 20.57 54.54 292.17Number of Sells 4687 36.51 39.85 1.00 10.92 18.99 50.20 286.43Buy-Sell Imbalance 4687 5.12% 7.66% -34.74% 1.04% 5.06% 9.23% 37.80%
Panel B: 2007-2017 TAQ
Retail TradesNumber of Buys 2768 576.50 200.01 79.10 496.41 599.13 708.18 1555.10Number of Sells 2768 547.66 185.72 83.38 474.09 569.83 664.95 1523.12Buy-Sell Imbalance 2768 2.25% 3.95% -18.05% 0.07% 2.63% 4.83% 17.57%
Each day we compute the number of small and large buyer and seller initiated trades for the ISSM/TAQ 1983-2000 data and the number of retail buys and sells for the 2007-2017 data. We then compute the buy-sell imbalance, BSI = (#Buys - #Sells) / (#Buys + #Sells), for that day. Panel A reports summary statistics based calculated from daily observations from 1983 to 2000. Small trades are for $5,000 or less (1991 dollars) or less; large trades are for $50,000 or more (1991 dollars). Panel B reports summary statistics calculated from daily observations from 2007-2017.
Panel A: Daily Data
Panel B: Weekly Data
Figure 1: Buy-Sell Imbalance based on Abnormal Volume Sorts
Panel A: Daily Data
Panel B: Weekly Data
Figure 2: Buy-Sell Imbalance based on Return Sorts
ISSM/TAQ Signed Trades, 1983-2000 Retail Trades, 2007-2017
In each week (Panel A) or day (panel B), stocks are sorted into 20 groups based on their returns in the preceding period. During the next period after the sort, we calculate the buy-sell imbalance as the number of buys less the number of sells divided by the sum of buys and sells across all stocks within the return sort. The two figures in the left column are based on large (blue) or small (red) trades based on ISSM/TAQ data from 1983-2000. The two figures on the right column are based on the identification of retail trades in the TAQ data from 2007-2017. Time series mean Buy-Sell Imbalances are graphed together with 95% confidence intervals.
Retail Trades, 2007-2017
In each week (Panel A) or day (panel B), stocks are sorted into 20 groups based on their abnormal trading volume during the period. During the same period as the sort, we calculate the buy-sell imbalance as the number of buys less the number of sells divided by the sum of buys and sells across all stocks within the abnormal volume sort. The two figures in the left column are based on large (blue) or small (red) trades based on ISSM/TAQ data from 1983-2000. The two figures on the right column are based on the identification of retail trades in the TAQ data from 2007-2017. Time series mean Buy-Sell Imbalances are graphed together with 95% confidence intervals.
ISSM/TAQ Signed Trades, 1983-2000
-30%
-20%
-10%
0%
10%
20%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Abnormal Volume Vingtile
-30%
-20%
-10%
0%
10%
20%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Abnormal Volume Vingtile
Large Trades
Small Trades
Large Trades
Small Trades
Retail Trades
Retail Trades
-30%
-20%
-10%
0%
10%
20%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Abnormal Volume Vingtile
Large Trades
Small Trades
-30%
-20%
-10%
0%
10%
20%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Abnormal Volume Vingtile
Retail Trades
-3%
0%
3%
6%
9%
12%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Return Vingtile
-3%
0%
3%
6%
9%
12%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Return Vingtile
Large Trades
Small Trades
Large Trades
Small Trades
Retail Trades
Retail Trades
-3%
0%
3%
6%
9%
12%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Return Vingtile
Large Trades
Small Trades -3%
0%
3%
6%
9%
12%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Return Vingtile
Retail Trades
Panel A: Daily Data Panel B: Weekly Data
Figure 3: Buy-Sell Imbalance based on Abnormal News Sorts
ISSM/TAQ Signed Trades, 1983-2000
In each week (Panel A) or day (panel B), stocks are sorted into 5 groups based on their abnormal news in the period. The abnormal news measure is computed as the number of news stories in the current week (or day) divided by the average number of weekly (or daily) news stories during the reference period from week -54 to week -5 (or from day -270 to day -21). A minimum of 10 weeks (or days) with news during the reference period is required. Stocks with no news in the current week or day are assigned to bin 0. Remaining stocks are sorted into quartiles on abnormal news measure, and are assgined to bins 2 (low abnormal news), 3, 4, and 5 (high abnormal news). In the same period as the sort, we calculate the buy-sell imbalance as the number of buys less the number of sells divided by the sum of buys and sells across all stocks within the abnormal news sort. The two figures are based on large (blue) or small (red) trades based on ISSM/TAQ data from 1983-2000. Time series mean Buy-Sell Imbalances are graphed together with 95% confidence intervals.
0%
2%
4%
6%
8%
1 2 3 4 5
Buy
-Sel
l Im
bala
nce
Abnormal News Bin
Large Trades
Small Trades
Large Trades
Small Trades
0%
2%
4%
6%
8%
1 2 3 4 5
Buy
-Sel
l Im
bala
nce
Abnormal News Bin
Large Trades
Small Trades
Figure 4: Cumulative Number of Small Buys minus Small Sells for EntreMed
TAQ Signed Trades, 10/01/1997 - 12/31/1998
The cummulative number of small (large) buyer initiated trades in EntreMed minus small (large) seller initiated trades beginning Oct. 1, 1997 (left axis). Daily closing price of EntreMed (right axis).
0
12
24
36
48
60
72
84
96
108
120
-5,000
-4,000
-3,000
-2,000
-1,000
0
1,000
2,000
3,000
4,000
5,000
ClosingPrice
Cum
ulat
ive
num
ber o
f B
uys m
inus
Sel
ls
Large Trades
Small Trades
Price
November 28, 1997
May 4, 1998
Figure 5: Buy-Sell Imbalance on Dartboard Column Stocks
ISSM/TAQ Signed Trades, 1988-2000
The buy-sell imbalances, small and large, are reported on stocks by the Dartboard column of Wall Street Journal. The sample covers 145 columns during Oct. 4, 1988 and Dec. 14, 2000, which include 480 stocks under Pro Picks and 475 stocks under Dartboard Picks. A minimum of 10 trades per stock-day is required when computing buy-sell imbalances. Buy-sell imbalances are reported on each day of the event window [t-10, t+10]. Mean buy-sell imbalances and 95% confidence intervals are graphed for each event-time day.
-10%
0%
10%
20%
30%
40%
t-10 t-9 t-8 t-7 t-6 t-5 t-4 t-3 t-2 t-1 t t+1 t+2 t+3 t+4 t+5 t+6 t+7 t+8 t+9 t+10
Buy
-Sel
l Im
bala
nce
Event Time Date
Large Trades
Small Trades
-10%
0%
10%
20%
30%
40%
t-10 t-9 t-8 t-7 t-6 t-5 t-4 t-3 t-2 t-1 t t+1 t+2 t+3 t+4 t+5 t+6 t+7 t+8 t+9 t+10
Buy
-Sel
l Im
bala
nce
Event Time Date
DartboardPicks
Small Trades
Large Trades
ProPicks
Figure 6: Buy-Sell Imbalances based on Mad Money covered Stocks
TAQ Retail Trades, 2007-2017
The buy-sell imbalances, small and large, are reported on stocks by the CNBC's TV show Mad Money. The sample covers 42,439 stock recommendations from 2007 to 2017 (only including US domestic stocks with share code 10 or 11). A minimum of 10 trades per stock-day is required when computing buy-sell imbalances. Mean Buy-sell imbalances and 95% confidence intervals are reported for each day of the event window [t-10, t+10].
0%
1%
2%
3%
4%
t-10 t-9 t-8 t-7 t-6 t-5 t-4 t-3 t-2 t-1 t t+1 t+2 t+3 t+4 t+5 t+6 t+7 t+8 t+9 t+10
Buy
-Sel
l Im
bala
nce
Date
Retail Trades
Figure 7: Buy-Sell Imbalances based on Abnormal Google Search Volume
TAQ Retail Trades, 2007-2017
A week starts on Sunday and ends on Saturday. Weekly search volume information is collected from Google Trends. In each week, the stocks (excluding the bottom 10% stocks in the previous year end's market capitalizations) are sorted into 20 vingtiles based upon abnormal search volume index (ASVI). ASVI is defined as the logarithm of a stock's current week's search volume index minus the median of the logarithm of each of the preceding 8 weeks' search volume indicies. During the same period as the sort, we calculate the buy-sell imbalance as the number of buys less the number of sells divided by the sum of buys and sells across all stocks within the abnormal search volume sort. The buy-sell imbalances are based on the identification of retail trades in the TAQ data from 2007-2017. Time series mean Buy-Sell Imbalances are graphed together with 95% confidence intervals.
-2%
-1%
0%
1%
2%
3%
4%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Buy
-Sel
l Im
bala
nce
Abnormal Search Volume Vingtile
Retail Trades